Risk of severe upper gastrointestinal complications among oral bisphosphonate users

Abstract

Background: Oral bisphosphonates (BPs) are the primary agents for the treatment of osteoporosis. Although BPs are generally well tolerated, serious gastrointestinal adverse events have been observed.

Aim: To assess the risk of severe upper gastrointestinal complications (UGIC) among BP users by means of a large study based on a network of Italian healthcare utilization databases.

Methods: A nested case-control study was carried out by including 110,220 patients aged 45 years or older who, from 2003 until 2005, were treated with oral BPs. Cases were the 862 patients who experienced the outcome (hospitalization for UGIC) until 2007. Up to 20 controls were randomly selected for each case. Conditional logistic regression model was used to estimate odds ratio (OR) associated with current use of BPs after adjusting for several covariates. A set of sensitivity analyses was performed in order to account for sources of systematic uncertainty.

Results: The adjusted OR for current use of BPs with respect to past use was 0.94 (95% CI 0.81 to 1.08). There was no evidence that this risk changed either with BP type and regimen, or concurrent use of other drugs or previous hospitalizations.

Conclusions: No evidence was found that current use of BPs increases the risk of severe upper gastrointestinal complications compared to past use.

Figure 1. Study flow-diagram. AIFA-BEST project, Italy, 2003–2007.

BPs: Bisphosphonates.

Figure 2. Adjusted odds ratios (and 95% confidence intervals) of upper gastrointestinal complications associated with current use of bisphosphonates within various patient subgroups. AIFA-BEST project, Italy, 2003–2007.

Odds ratios estimated with conditional logistic regression model. Estimates concerning main analysis were unadjusted and adjusted for use of other drugs and for the number hospitalizations in the 60-day period prior the index date. Estimates concerning subgroup analysis were obtained by including the interaction terms combining the effect of current use of BPs together with BPs type and regimen dispensed during the current period, concurrent use of other drugs and number hospitalizations in the 60-day period prior the index date. P-values concern comparison of BPs effect across patient subgroups or along increasing number of hospitalizations. BPs: Bisphosphonates.

Figure 3. Influences of diagnostic criteria for upper gastrointestinal complications (panel A), length of time-window for current use of bisphosphonates (panel B), and of controlling for protopathic bias (panel C) on the observed odds ratio for upper gastrointestinal complications associated with current use of bisphosphonates. AIFA-BEST project, Italy, 2003–2007.

Estimates are adjusted for use of other drugs and number of hospitalizations in the 60-day period prior the index date. Details for diagnostic criteria are reported in Appendix S1. For an explanation of methods for controlling protopathic bias see the “Sensitivity analysis”, subsection of the “Methods” section. BPs: Bisphosphonates.

Figure 4. Modelled influence of the inclusion of prevalent BPs users on the true association between current BPs and UGIC risk. AIFA-BEST project, Italy, 2003–2007.

The graph indicates the trend of the true effect of BPs current use on the UGIC risk (e.g. the odds ratio which we would have observed if only incident users were included) according to different values of the BPs – UGIC association among prevalent users. For an explanation see the “Sensitivity analysis”, subsection of the “Methods” section. BPs: Bisphosphonates. UGIC: Upper gastrointestinal complication.

Figure 5. Modelled influence of a hypothetical confounder unaccounted for in the adjustments already performed in the main analysis according with the direction of its effect on the outcome (i.e. positive and negative associations as reported in boxes A and B, respectively), and with its prevalence in the study population (p). AIFA-BEST project, Italy, 2003–2007.

The graphs indicate what combinations of confounder – UGIC and confounder – current BPs exposure would be required to make statistically significant the observed association between current use of BPs and hospitalization for UGIC. For an explanation see the “Sensitivity analysis”, subsection of the “Methods” section. BPs: Bisphosphonates. UGIC: Upper gastrointestinal complication.